MaterialandMethodsS1.-Plos.docVIP

Material and Methods S1 A Diffusion MRI acquisition Five data sets are acquired at 3T with an Achieva (Philips, Einthoven, The Netherlands) MRI scanner using a diffusion weighted spin echo EPI technique [1, 2]. The timing parameters of the pulse sequence are TR/TE/Δ/δ = 4200/89/43.5/32.5 ms, maximum diffusion gradient intensity is 80 mT/m, yielding a maximal b value of 9000 s/mm2 [3]. The matrix size is 112x112 and the slice number is 36, with a voxel size of 2x2x3 mm3. Diffusion imaging is performed according to the classical Diffusion Spectrum Imaging (DSI) scheme as described in [3]. Briefly, diffusion-weighted images covering the whole brain are acquired for 129 different values of diffusion sensitizing gradient intensity and direction (i.e., different q-vectors) [4], comprising in q-space the points of a cubic lattice within a hemisphere of 4 lattice units in radius. We take with integers and , and, denoting the unit diffusion sensitizing gradient vectors in the three respective coordinate directions. Next, we process these 129 images as follows. First, we reconstruct the 3D diffusion function, or Probability Density Function (PDF) at each brain location by taking the discrete 3D Fourier transform of the signal modulus sampled in q-space. The signal is pre-multiplied by a Hanning window before Fourier transformation in order to ensure a smooth attenuation of the signal at high values. With this procedure and the above parameters the PDF is sampled over an isotropic 3-dimensional field of view of 100?m with a nominal isotropic resolution of 10?m. The result, called a diffusion map, is an image that associates a 3D diffusion function with every brain position voxel. From this map, at each voxel, we compute an Orientation Density Function (ODF) ?(u), by projection of the PDF in the radial direction. For u a 3D vector with |u| = 1, we define: (1) where p(.) is the 3D PDF, ? is the radius, ? 2d? is the 3D volume element and the integral is evaluated as a discret